Loading...
Please wait, while we are loading the content...
Similar Documents
Broadband network traffic modeling, management and fast simulation based on Ã-stable self-similar processes
| Content Provider | Semantic Scholar |
|---|---|
| Author | Karasaridis, Anestis |
| Copyright Year | 1999 |
| Abstract | We propose a new moclel for network heavy-trüffic approxiniation. basetl on a-S table selfsimilar processes. riamely the skewecl Linear Fractional Stable Yoise. The model is long-range dependent and demonstrates more fiexibility than existing niodeis in fitting different levels of biirstiness and dependence iri the data. Nonetlieless. it is parsimonious in the number of parameters. which have a direct physical meaning. The marginal distribution of the model is a-Stable. and therefore the Generalized Central Limit Theorem can be applied to provide a physical interpretation on how aggregate effects in traffic appear as a superposition of traffic from independent sources. ÇVe present an algorithm for the estimation of the model parameters. which is based on properties of the Totally Skewed a-S table distribution. We investigate the implications of our proposed rnodeling on the estimation of bandwidth allocation and admission control of bursty, long-range dependent sources. Analytic formulas for the overfiow probability bounds of a constant service rate buffer are derived, and they are used to provicle boonds of' the required bandwidth allocation of a-Stable self-similar sources. for which the general theory of effective bandmidths does not a p p l Extensive simulations are presented. where the new model is fitted to bursty Ethernet and variable-bit-rate video data. Furthermore. new measurernents of aggregate web and webcasting traffic are introduced dong with traffic generated by the fitted new model. Qoeueing simtilations witli real traffiç support our ürialytical results for the overflow probebility. The espression of the iippcr bandwiclth allocation bound is simple and allows quick computation of aclmissible regions for multiplesed liomogeneous or heterogeneous sources entering the network. Oiir analyt ical resiil ts for the equivalent rate boiinds are verified wi t h extendeci simulation stiidics mit h real and rnodel generated traffic. In ttic context of prrdicting rare cvents, such as ceIl Iosses in networks. more eficiently. we present ;i ncw niethoci for fast sirniilation of rare events which follow an O-Stable distribiition. Tlio proposeci nietliod provides considerable reduction in the number of samples requiretl for acciiratc estimation. coniparcd to a Monte-Carlo simulation. Furthermore. the variance of t bc cstirtiator cari be t hree orders of magnitude sniallcr t han the traditional bleari-Translation niethod iri the Iniportancc Sampling framework. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | https://tspace.library.utoronto.ca/bitstream/1807/12842/1/NQ45688.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
